Clinical, Laboratory and Tomographic characteristics associated with obesity and BMI at Hospital admission in adult patients with COVID-19: a cross-sectional study

Introduction: The pathological status of obesity can influence COVID-19 from its initial clinical presentation, therefore, the identification of clinical and laboratory parameters most affected in the presence of obesity can contribute to improving the treatment of the disease. Objective: To identify the clinical, laboratory, and tomographic characteristics associated with obesity and BMI at t hospital admission in adult patients with COVID-19. Methods: This is a cross-sectional observational study with a total of 315 participants with COVID-19 confirmed by rt-PCR. The participants were divided into non-Obese (n=203) and Obese (n=112). Physical examinations, laboratory tests, and computed tomography of the chest were performed during the first 2 days of hospitalization. Results: Patients with obesity were younger, and they had higher systolic and diastolic blood pressure, higher frequency of alcoholism, fever, cough, and headache, higher ALT, LDH, and red blood cell count (RBC), hemoglobin, hematocrit, and percentage of lymphocytes. Also, they presented a lower value of leukocyte count and Neutrophil/Lymphocyte Ratio (RNL). The parameters positively correlated with BMI were alcoholism, systolic and diastolic blood pressure, fever, cough, sore throat, number of symptoms, ALT in men, LDH, magnesium, RBC, hemoglobin, hematocrit, and percentage of lymphocytes. The parameters negatively correlated with the BMI were: age and RNL. Conclusion: Several parameters were associated with obesity at hospital admission, revealing better than expected results. However, these results should be interpreted with great caution, as there may be some influence of a phenomenon called the Obesity Paradox that can distort the severity and prognosis of the patient


INTRODUCTION
Obesity started to be considered a disease for the first time in 2013, by the American Medical Association, and since then it was realized that it was a disease with high prevalence in many countries, reaching pandemic proportions 1 . Worldwide, in 2016, 39% of adults were overweight and 13% were obese 2 . In Brazil, in 2018, 55.4% were overweight and 20.3% were obese 3 . The problem of the high and increasing prevalence of obesity in the world is due to the risk for other non-communicable diseases and chronic conditions, including type 2 diabetes mellitus, non-alcoholic fatty liver disease, systemic arterial hypertension (SAH), myocardial infarction, stroke, dementia, osteoarthritis, obstructive sleep apnea and various types of cancer 4 .
In addition to the influence on non-communicable diseases, obesity has also been considered an important factor associated with complications of the disease caused by the new coronavirus (COVID- 19), with a negative influence on the general hospitalization rate, hospitalization in the intensive care unit, disease severity, and death, also contributing to the development of more virulent strains 5 . Among the possible justifications for the worse prognosis of patients with obesity and COVID-19 are the chronic and intrinsic changes in obesity, mainly metabolic changes and in the endocrine, nervous, immune, cardiovascular, respiratory, and hematological systems 6 . In addition, there is still no specific guidance on the best approach for a patient with obesity and COVID-19 7 .
The pathological status of obesity can influence COVID-19 from its initial clinical presentation, therefore, the identification of clinical and laboratory parameters

Data collection
The State Field Hospital of Piaui was planned since its creation to be a health institution for research purposes. Therefore, the data collection instruments used were specific to this work, where the hospital staff was previously trained. All data were collected in the first contact with the patient, with complementary exams performed within 48 hours after admission.
The following clinical data were self-reported and recorded in standardized electronic forms pre-defined by the research group: Age, sex, days of symptoms until hospitalization, weight, height, diseases, signs, symptoms, presence of smoking, and alcohol consumption. Smoking was defined as the continuous use of cigarettes and similar products in the last 6 months. Alcoholism was defined as the weekly ingestion of any alcoholic product in moderate to large doses in the last 6 months, according to the patient's perception. Vital signs were measured by the nursing team using standardized hospital instruments and recorded on specific forms: axillary temperature, heart rate, respiratory rate, systolic and diastolic blood pressure, oxygen saturation, and capillary blood glucose. Laboratory tests and computed tomography (CT) of the chest were under the responsibility of a single company experienced in the market and all the information was transferred in full to a database with double entry, to avoid typing errors, with subsequent review by a team of experienced doctors. Laboratory data collected were ALT, AST, creatinine, LDH, lactate, ferritin, CRP, sodium, potassium, magnesium, red cell count, hemoglobin, hematocrit, white cell count, and platelet count. Chest CT data were collected from a specific form with result categories and ground-glass opacification levels in case there was this change.

Statistical analysis
After applying the normality test, the quantitative variables of the obese and nonobese groups were expressed as a median and interquartile range, and comparisons between groups were performed using the Mann-Whitney U test. Qualitative variables were presented as absolute frequency and percentages, and the associations were tested using Pearson's Chi-Square Test. To test the correlation of variables with the body mass index, the Spearman correlation coefficient (ρ) was used for the quantitative variables, and the point-biserial correlation coefficient (r) for the dichotomous qualitative nominal variables. No imputation was made for missing data. The level of significance was set at p<0.05. The study data were processed using the Statistical Package for the Social Sciences software, version 27 (IBM Inc., Chicago, IL, USA).

RESULTS
The sample consisted of 313 patients, 35 Regarding general clinical parameters, patients with obesity and COVID-19 were younger (p<0.01), had a higher frequency of drinking (p<0.01), had higher systolic blood pressure (SBP) (p<0.05) and had higher diastolic blood pressure (DBP) (p<0.01).
Regarding laboratory parameters (Table 3) p<0.05). Ground-glass opacification showed no difference between the obese and nonobese groups and also did not correlate with BMI (Table 3).

DISCUSSION
In this study, obese patients with SARS-CoV-2 infection were younger than those without obesity. Excess weight has been considered an independent risk factor for hospitalization. Although young adults are considered to be at lower risk for hospitalization and disease severity, they became more prevalent when obese 5 . Furthermore, we found an inverse relationship between BMI and age that was also found in similar studies 9-12 . COVID-19 was not consistent with Brazilian epidemiology, where obesity is more prevalent in older adults 3 . An important study with data from Brazil pointed out that patients with obesity who died from COVID-19 are younger than patients without obesity. The unfavorable interaction between pathogen, host, and environment in the younger adult with obesity is not yet elucidated 13 .
The association between alcoholism and obesity found in this study can be justified by some complementary hypotheses: 1) alcohol consumption is usually higher among people with obesity; 2) there was an increase in alcohol consumption during confinement resulting from the pandemic and 3) concomitance between obesity and alcoholism causes overexpression of type I angiotensin II receptors, impairs the immune system and leaves the patient more likely to develop symptomatic COVID-19 14,15 .
The median systolic and diastolic blood pressure of the obese and non-obese groups fit as stage 1 systemic arterial hypertension, according to the American College of Cardiology and the American Heart Association 16 , meaning that at least 50% of all patients had SAH. When observing the prevalence of self-reported SAH, we found an agreement in the results, since 51% of the non-obese group and 58.6% of the obese group reported having the disorder. The study by Ritter et al. 17 listed a series of biological mechanisms that occur in obesity and influence the development of SAH: systemic inflammation, endothelial damage, higher plasma concentrations of renin, angiotensinogen, angiotensin II conversation enzyme, angiotensin II and aldosterone, with greater activation of the renin-angiotensin-aldosterone system, among others.
Taing's cross-sectional study 18 with the general population of India found a considerable positive correlation between BMI and systolic and diastolic blood pressure, as observed in this study. Therefore, obesity is an independent factor for the increase in blood pressure and its influence is proportional to the increase in BMI. Excess adipose tissue may be responsible for a more symptomatic presentation of COVID-19 7,17 . We found in this study a higher prevalence of fever, cough, and headache in patients with obesity, in addition, fever, cough, sore throat and the number of self-reported symptoms were directly proportional to the increase in BMI. Although fever was self-reported by more than 60% of patients, the temperature was normal in more than 75%, probably due to self-medication. The two symptoms that were more frequent in patients with obesity were also the two most frequent in the general population 19 . The prospective cohort study by Nilles et al. 20 , which collected BMI information, sociodemographic data, comorbidities, and symptoms from patients with COVID-19, showed that obesity was associated with an increase in the reporting of symptoms, mainly selfreported fever, but not measured. Another interesting finding from the work of Nilles et al. 20 was that the clinical condition is strongly influenced by obesity among younger age groups. The scientific literature says that patients with obesity have more expressive clinical conditions since the onset of symptoms and this is a possible explanation to justify the greater risk of hospitalization 5 . What can justify the most symptomatic illness process of patients with obesity are the following changes resulting from obesity: chronic inflammatory state, diffuse cardiovascular involvement, reduced respiratory capacity, impaired immune response, overexpression of the angiotensin conversation enzyme (ACE2), and greater virulence 17,21 .
The patient with obesity has several altered laboratory parameters due to the overlapping pathophysiological molecular mechanisms of obesity and COVID-19 17 . In the men in this study, we observed that ALT is proportional to the increase in BMI, which may reveal a predisposition for this association in males. The association between obesity and higher ALT may be before the disease caused by SARS-CoV-2 since patients with obesity have a higher risk of having an abnormal liver function, or it may be a consequence of the acute inflammation of the current disease 22,23 .
The acute phase inflammatory markers are commonly altered in patients with obesity and COVID-19, and LDH is one of the most frequently elevated acute phase inflammatory factors and has the potential to predict prognosis 24  Serum magnesium levels were higher in patients with obesity and also showed a positive linear correlation with BMI. Hypermagnesemia is common in patients with obesity in general 27 , however, when associated with insulin resistance of diabetes or metabolic syndrome, there is serum hypomagnesia, due to intracellular changes 28 . As the postprandial capillary glycemia measured in the two groups did not show a statistical difference between them, as well as the absolute frequency of self-reported diabetes, we can assume that the glucose metabolism also influenced the magnesium metabolism of both groups and the final result was a higher level of magnesium in patients with obesity, but within normal limits.
Studies that associate obesity and hematological changes in humans are scarce 29 .
No articles were found on the association between hematological changes and patients with obesity and COVID-19. Therefore, this work aims to create hypotheses that justify the changes found in the complete blood count.
The red blood cell, hemoglobin, and hematocrit values were higher in patients with obesity. Obesity and increased waist circumference increase the production of red cells by several different mechanisms 29,30 . In general, inflammation is associated with the suppression of erythropoiesis, but even with chronic inflammation caused by obesity, there is an increase in erythropoiesis in patients with obesity. The study by Vuong et al. 30 listed mechanisms that explain the increase in red blood cells, hematocrit, and hemoglobin in patients with obesity: increased insulin and insulin-like growth factors in people with metabolic syndrome; chronic hypoxia resulting from the pathophysiology of obesity, which can be aggravated by the common association with obstructive apnea; and, finally, leptin and adiponectin levels may be related to erythrocyte metabolism. In patients with non-severe COVID-19 regardless of BMI, there is an increase in erythrocytes, an increase in hemoglobin, and a reduction in RDW, so we can conclude that our patients with obesity and COVID-19 had less severity, based on the blood count parameters 31 .
Regarding the leukogram, obesity causes a series of pathophysiological changes that result in the alteration of leukocytes. In people with obesity, adipose tissue has hypertrophied adipocytes with lipid engorgement, cell stress, and frequent apoptosis that shown that patients with obesity normally had higher values of total leukocytes, neutrophils, and lymphocytes than patients without obesity, these elevations being proportional to the BMI 29,33 . There is still no consensus on the influence of obesity on the immune system in response to an infection. According to Green and Beck's review 34 , most authors argue that obesity impairs the immune response, while some argue that there are protective mechanisms involved, depending on the acquired infectious disease. People with obesity and infectious diseases generally have high leukocyte and neutrophil values 17 , and a reduction in the production and circulation of T lymphocytes 34 . About COVID-19, it was found that patients with obesity with more severe disease had an increase in the total leukocyte count, a reduction in lymphocytes, and an increase in the neutrophil/lymphocyte ratio (RNL) and a lower percentage of basophils, monocytes, and eosinophils 35 .
The absolute values of total leukocytes and neutrophils in people with obesity found in this work were curiously lower. The only parameter of the comparatively high leukogram in this study was the relative value of leukocytes, which also showed an upward trend in the absolute value, but remains lower than the lower limit of normality. Therefore, all parameters presented an increase or decrease opposite to that expected in patients with obesity and infection. The unexpected change in leukocytes in patients with obesity and COVID-19 may be a peculiarity related to SARS-CoV-2. The meta-analysis by Zeng et al. 36 evaluated the influence of the leukogram parameters on the severity of COVID-19. Their results showed that higher leukocyte and neutrophil counts are related to greater disease severity, as well as higher neutrophil/lymphocyte ratio (NLR) and lower  36 . With that, we can conclude that the leukogram values of patients with obesity were associated with less severity.
The patients with obesity observed in this study did not present different tomographic findings of the chest from the patients without obesity, despite high LDH values, chronic inflammation, and impaired respiratory physiology characteristic of obesity 17 . The tomographic presentation of patients with COVID-19 and what can influence them is not fully elucidated. Some authors suggest that there is an association between greater pulmonary impairment on chest CT and inflammatory markers, however, the relationship with the severity and prognosis of the disease is not entirely clear 26,37 . As no reliable correlation has yet been found between chest CT and COVID-19 severity/prognosis, the WHO recommends that this imaging test be performed only as a diagnostic resource where there is no laboratory test and to assist in making decisions about the location of hospitalization, whether in the infirmary or the intensive care unit and not as a parameter to weigh prognosis 38 .
This study has some limitations that should be highlighted, with emphasis on 1) In this study, we observed that all laboratory parameters at the admission of patients with obesity indicated less or equal severity to those without obesity. This observation is consistent with the initial risk classification, which identified a similar percentage of severe cases between groups. It has been suggested that a phenomenon called the "Obesity Paradox" refers to the lower severity of some respiratory, cardiovascular, and renal diseases in patients with obesity in the short and medium term 39 .
Regarding COVID-19, the current literature shows that the paradox does not apply, as obesity worsens the overall prognosis 5,6 . However, the studies that reached this conclusion evaluated the risk of hospitalization, hospitalization in an intensive care, need for intubation, and death. This work evaluated something different: the clinical, laboratory, and tomographic findings in the first days of hospitalization, which configures an approach that has not yet been explored. In addition, there is a possibility of involvement, at least in part, of the Obesity Paradox in the results of the less severe laboratory parameters found in the first days of hospitalization. However, as mentioned earlier, this work has a limitation related to the statistical power of the sample, whose size can influence the statistics found. Thus, further studies are needed to confirm this possibility.
Finally, we concluded that patients with obesity and COVID-19 at hospital admission were younger, had a higher frequency of alcoholism, fever, cough, and headache, and had higher levels of DPB, SPB, ALT, LDH, red blood cells, hemoglobin, hematocrit and relative lymphocytes, and lower values of leukocytes, neutrophils, and NLR. In addition, BMI was positively correlated with alcoholism, DBP, SBP, fever, cough, sore throat, number of symptoms, ALT in men, LDH, magnesium, red blood cells, hemoglobin, hematocrit, and relative lymphocytes; and negative correlation with age and NLR. In this way, it is concluded that several parameters were associated with obesity at hospital admission, revealing better than expected results in the patients with obesity and COVID-19. However, these results should be interpreted with great caution, as there may be some influence of a phenomenon called the Obesity Paradox that can distort the severity and prognosis of the patient. The comparison data were presented using the median and interquartile range and compared using the Mann-Whitney U test or presented using the absolute frequency and percentage and compared using Pearson's chi-square test. The correlation data were presented as a point-biserial correlation coefficient (r) or Spearman's (ρ), as appropriate. For all tests, results were considered statistically significant when p<0.05. # p<0.10; * p< 0.05; ** p< .01; SAH: systemic arterial hypertension; n: total number of observations per group.